The influence of text length and prior knowledge on
the testing effect in meaningful learning
Sandra Wetzels and Liesbeth Kester
L&C-plenary
June 14, 2012
Theoretical background:
The testing effect
 Testing effect often studied in verbal learning tradition
 Testing effect paradigm:
study phase – initial test/restudy – delayed final test
 Performance on delayed final test boosted most by prior testing
 Tests are powerful learning tools
Theoretical background:
Rote versus meaningful learning
 Research focused on testing isolated facts (rote learning)
 Limits relevance of findings for educational practice, which
focuses on meaningful learning
 Research on testing effect for more educationally relevant
ideational units is scarce
Theoretical background:
Testing effect in meaningful learning
Karpicke and Blunt (2011)
Kester, Firssova, Gorissen,
Wetzels, and Kirschner (2011)
276-word science text on sea otters
850-word texts on important
advances in history of science
Study phase – restudy/free recall –
delayed final short-answer test
Study phase – restudy/free
recall/short-answer test – delayed
final short-answer test
Testing effect for verbatim questions Testing effect for closed questions
(rote learning)
(rote learning)
Testing effect for inference
questions (meaningful learning)
No testing effect for open questions
(meaningful learning)
Theoretical background:
The influence of text length
 Methodological difference Karpicke and Blunt (2011) and Kester
et al. (2011): text length (276 words versus 850 words)
 Text length/number of ideational units may influence testing
effect in meaningful learning
 Build semantic network from text that contains important
concepts and interrelations
 The longer a text, the more difficult it is to build a semantic
network during the study phase
Theoretical background:
The influence of prior knowledge
 Prior knowledge plays pivotal role in meaningful learning
 Desirable difficulties framework (Bjork, 1994, 1998):
Difficult but successful processing is more beneficial for
retention than difficult but unsuccessful processing
 Retrieval effort hypothesis (Pyc & Rawson, 2009):
if retrieval is successful, more difficult retrieval is better for
retention than less difficult retrieval
 Successfulness and difficulty retrieval attempt is influenced by
learners’ level of prior knowledge
 Prior knowledge might also influence learners’ ability to build
semantic network from text
Research question
The occurrence and magnitude of the testing effect in meaningful
learning is influenced by (1) text length and more specifically,
the number of ideational units, and (2) prior knowledge
Hypotheses
Hypothesis 1
For novice learners, taking a free-recall test is expected to result
in higher performance on the verbatim questions of a delayed
short-answer test taken after one week than restudying
independent of text length
Hypothesis 2
For novice learners, taking a free-recall test is expected to result
in higher performance on the inference questions of a delayed
short-answer test taken after one week than restudying for the
short text
Hypothesis 3
For novice learners, taking a free-recall test is not expected to
have any beneficial effects on performance on inference
questions of a delayed short-answer test taken after one week
over and above restudying for the long text
Hypothesis 4
For more advanced learners, taking a free-recall test is expected
to yield higher performance on both verbatim and inference
questions of a delayed short-answer test taken after one week
compared to restudying for both the short and the long text
Hypothesis 5
Restudying is expected to yield the highest delayed short-answer
test performance after 5 minutes independent of text length or
learners’ level of prior knowledge
Method:
Participants and materials
Participants
Fourth-year students in secondary education (Study 1a) and
students from higher education (Study 1b)
Materials
- Two texts about important advances in history of science; one
text is short (300 words) and one text is long (900 words)
- Multiple-choice prior knowledge test
- Initial free-recall test
- Delayed final short-answer test
Method: Design
Design
2 x 2 x 2 mixed factorial design with the factors:
- Text length/number of ideational units (short/low versus
long/high)
- Learning strategy (free-recall test versus restudy)
- Retention interval (five minutes versus 1 week)
Dependent variables:
- Performance on verbatim questions short-answer test
- Performance on inference questions short-answer test
- Transfer of retrieval from initial test to delayed final test
Method: Procedure
Procedure
- Prior knowledge test (one week before experimental session)
- Study text 1 (5 or 15 minutes)
- Sudoku puzzle (2 minutes)
- Restudy or free-recall test (10 minutes)
- Sudoku puzzle (2 minutes)
- Study text 2 (5 or 15 minutes)
- Sudoku puzzle (2 minutes)
- Restudy or free-recall test (10 minutes)
- Sudoku puzzle (5 minutes)
- Delayed final short-answer test (after 5 minutes or 1 week)
[email protected]